Temporal trends and mortality associated with difficult-to-treat resistance in bloodstream gram-negative bacilli: a retrospective cohort study in Southern Saudi Arabia

Abstract

Introduction Antimicrobial resistance among bloodstream gram-negative bacilli is a major driver of infectious-disease mortality, yet broad resistance labels can obscure clinically important heterogeneity. We aimed to characterise long-term organism and resistance trends, including difficult-to-treat resistance (DTR), among bloodstream gram-negative bacilli and to estimate their association with 30-day in-hospital mortality. Methods We conducted a retrospective cohort study of hospitalized patients aged 12 years or older with blood cultures yielding Enterobacterales or non-fermenting gram-negative bacilli (GNB) at a 500-bed tertiary hospital in Southern Saudi Arabia from 1 January 2013 to 27 August 2024. Duplicates within 14 days were excluded; trends were analyzed with segmented binomial regression and mortality with modified Poisson regression adjusted for age, sex, intensive-care status, infection source, healthcare-associated infection, COVID-19 era, calendar year and Elixhauser comorbidity score. Results Among 2027 isolate-episodes from 1695 patients, 1487 (73.4%) were Enterobacterales and 540 (26.6%) non-fermenting GNB. The organism mix was broadly stable, but Klebsiella pneumoniae was dominant (37.6%) and increased (average annual percentage change (AAPC) +4.8%). Fluoroquinolone and extended-spectrum cephalosporin resistance declined among Enterobacterales, whereas carbapenem resistance (46.9%) and DTR (36.0%) rose, most sharply among Enterobacterales. Thirty-day in-hospital mortality was 36.5%. DTR was independently associated with higher mortality among Enterobacterales (adjusted risk ratio 1.35; 95% CI 1.15 to 1.59) and non-fermenting GNB (1.31; 1.09 to 1.57), with a graded increase as active first-line agents were lost. Discussion In this surveillance cohort, escalating carbapenem resistance and DTR, rather than a shifting organism mix, were associated with worse outcomes, supporting surveillance based on treatment-option phenotypes and prioritization of stewardship and infection-prevention efforts against high-risk GNB.